fix(autoplan): restore point and phase artifacts live in the project's .gstack/tmp/autoplan/

Skill hooks do not run inside subagents, so a reviewer launched after a
background completion notice raised a card for its own input (native-prompt.md
under the state root). Step 1 now puts RESTORE_PATH in the repository's
git-excluded .gstack/tmp/autoplan/ (added to .git/info/exclude, mode 700), and
the snapshot tool writes every phase artifact beside it, so /autoplan and its
reviewers read them from the working directory without a card. The guard's
bindings are unchanged: they anchor on the restore point's directory. Step 1
still prints the slug, which the old RESTORE_PATH carried.

A free test runs the generated Step 1 block twice from a repository
subdirectory: the restore point lands in <repo>/.gstack/tmp/autoplan/, the
exclude line is written once and git status stays clean. The autoplan parity
cap rises by the measured bytes (115,270 / 101,979 = 1.1303).

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
This commit is contained in:
garrytanandcapy-ai[bot] committed 2026-10-08 08:07:17 +00:00
1 parent 30f17d569a
commit 95fc35e585
4 files changed
+29 -9

No files matched your search

+18
View File
@@ -112,6 +112,24 @@ describe('/autoplan owned reads after the skill turn', () => {
expect(await runPublicationHook({ ...read(c.session, c.packet), tool_name: 'Bash', tool_input: { command: `cat ${c.packet}` } }, ROOT)).toEqual({});
});
test('Step 1 puts the restore point in the repository\'s git-excluded store, from a subdirectory too', () => {
const skill = fs.readFileSync(path.join(ROOT, 'autoplan/SKILL.md'), 'utf8');
const start = skill.indexOf('```bash\n', skill.indexOf('Fresh RESTORE_PATH')) + 8;
const block = skill.slice(start, skill.indexOf('\n```', start));
const home = scratch('autoplan-owned-home-'), repo = scratch('autoplan-owned-git-'), session = path.join(repo, 'packages', 'app');
fs.mkdirSync(path.join(home, '.claude', 'skills'), { recursive: true }); fs.mkdirSync(session, { recursive: true });
fs.symlinkSync(ROOT, path.join(home, '.claude', 'skills', 'gstack'), 'dir');
expect(spawnSync('git', ['init', '-q', '-b', 'main'], { cwd: repo, timeout: 10_000 }).status).toBe(0);
const run = () => spawnSync('bash', ['-c', block], { cwd: session, encoding: 'utf8', timeout: 15_000, env: { ...process.env, HOME: home } });
const first = run(), second = run();
expect(first.status).toBe(0); expect(second.status).toBe(0);
const restore = /^RESTORE_PATH=(.+)$/m.exec(first.stdout)?.[1] ?? '';
expect(path.dirname(restore)).toBe(path.join(repo, '.gstack', 'tmp', 'autoplan'));
expect(path.basename(restore)).toMatch(/-autoplan-restore-\d{8}-\d{6}\.md$/);
expect(fs.readFileSync(path.join(repo, '.git', 'info', 'exclude'), 'utf8').split('\n').filter(l => l === '/.gstack/tmp/')).toHaveLength(1);
expect(spawnSync('git', ['status', '--porcelain', '--untracked-files=all'], { cwd: repo, encoding: 'utf8', timeout: 10_000 }).stdout).toBe('');
});
test('a guard denial is never turned into an approval', async () => {
const cwd = scratch('autoplan-owned-deny-');
const output: any = await runPublicationHook(read(cwd, path.join(ROOT, 'autoplan/sections/design-phase.md')), ROOT);
+1 -1
View File
@@ -577,7 +577,7 @@ do not launch the downstream skill or open a browser.`,
maxSkeletonBytes: 70_000, // Phase-specific outside coverage, native fallback, and harness guard.
minUnionBytes: 85_000, // measured union 86,926
mustContain: ['6 Decision Principles', 'TASTE DECISION', 'USER CHALLENGE', 'consensus', 'Restore Point'],
maxSizeRatio: 1.1295, // Four validated outside invocations replace raw CLI calls; phases keep independent coverage. + Oct 7 wave B3: the Codex probe runs as a command (select-model capture lines, check-sandbox/probe-model/run-with-timeout subcommands) instead of being sourced; measured 114,297 / 101,979 = 1.1208 (2026-10-07). + Oct 7 wave CEO-28/DX-11: review prompts ask for a severity per finding and design proposals use the proposal gate; measured 114,533 / 101,979 = 1.1231 (2026-10-07). + Oct 7 wave guard (CEO-20): conditional `run_in_background` request plus the background-run recovery at each foreground dispatch site, and phase-close publishing each report with the `true autoplan-published <phase>` no-op before the next phase entry in a later message (CEO-1); measured 114,320 / 101,979 = 1.1210 (2026-10-07). Merged with the guard area: measured 115,149 / 101,979 = 1.1291 (2026-10-07).
maxSizeRatio: 1.1305, // Four validated outside invocations replace raw CLI calls; phases keep independent coverage. + Oct 7 wave B3: the Codex probe runs as a command (select-model capture lines, check-sandbox/probe-model/run-with-timeout subcommands) instead of being sourced; measured 114,297 / 101,979 = 1.1208 (2026-10-07). + Oct 7 wave CEO-28/DX-11: review prompts ask for a severity per finding and design proposals use the proposal gate; measured 114,533 / 101,979 = 1.1231 (2026-10-07). + Oct 7 wave guard (CEO-20): conditional `run_in_background` request plus the background-run recovery at each foreground dispatch site, and phase-close publishing each report with the `true autoplan-published <phase>` no-op before the next phase entry in a later message (CEO-1); measured 114,320 / 101,979 = 1.1210 (2026-10-07). Merged with the guard area: measured 115,149 / 101,979 = 1.1291 (2026-10-07). + Oct 8 wave (permission cards after a background reviewer): Step 1 puts RESTORE_PATH, and so every phase artifact, in the project's git-excluded `.gstack/tmp/autoplan/`, which Claude Code reads without a card after the skill turn ends; measured 115,270 / 101,979 = 1.1303 (2026-10-08).
},
spec: {
skill: 'spec',